instructor
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill's primary function is processing untrusted text (e.g., articles, reviews, user data) to extract structured fields, creating an attack surface for indirect prompt injection where malicious data could attempt to manipulate the model's output or extraction logic.
- Ingestion points: The skill ingests untrusted text in multiple examples across SKILL.md and references/examples.md.
- Boundary markers: The code snippets do not demonstrate the use of specific boundary markers or delimiters to isolate untrusted input within the prompt.
- Capability inventory: The skill utilizes remote LLM API capabilities (OpenAI and Anthropic); no local command execution, file system access, or additional network operations are present.
- Sanitization: Structural validation is performed on the output via Pydantic schemas, but the skill does not show specific sanitization for the input text.
- [DYNAMIC_EXECUTION]: The skill demonstrates the runtime creation of data validation models using Pydantic's 'create_model' function in SKILL.md, which is a form of dynamic logic generation.
Audit Metadata